TestCommander (autooptimisation) Trader's tool - page 8
You are missing trading opportunities:
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
Registration
Log in
You agree to website policy and terms of use
If you do not have an account, please register
Dear Igor!
When working with macro program "Complex" it turned out that in the report file of this macro program sometimes for some reason the profit results obtained by the macro program differ from the results obtained in manual mode with the same input data on the same tester! And there are no such profit results in the report file for optimization results (which is understandable). It is unclear how the "Complex" macro program obtained these results. Maybe we should check how the macro program reads the profit result from the tester report after the next test. The numbers of the report "Complex", have a correct format and appearance but are sometimes two or three times less or more than the real.
Regards Sergey.
This is most likely due to the order of the best parameters you have set (the order of screening)
The default sifting order in TestCommander is set:
1) Profit_Factor
2) Gross_Profit
3) Expected_Payoff
set these parameters in accordance with your requirements.
The sorting mechanism is described in details in the article 'Automated Optimization of Trading Robots in Live Trading'.
xeon, update expected soon?
It is well under way, but there is a lot of work to be done.
It's hard to say exactly when it's coming out, very roughly early next year.
It is well under way, but there is a lot of work to be done.
It's hard to say exactly when it will come out, very roughly early next year.
Eh Yo!
A couple of sheets ago it seemed like the end of that month, and now it's a year. Will the algorithms be completely rewritten or what?
What can we even expect from the new release?
Eh Yo!
A couple of sheets ago it seemed like the end of that month and now it's a year. Are the algorithms going to be completely rewritten or what?
What can we even expect from the new release?
Everything is being completely rewritten, even with the concept change, the graphical shell is changing.
The whole thing is being rewritten, even with a change of concept, the graphical shell is changing.
I see, so this is all for the long term.
Maybe the public could be asked to post the source code of the old development. Everyone would finish what they need.
As a last resort the restricted code.
As soon as new version will come out, with new endpoints and implementation, let it be commercial and then who needs the program will buy it.
I see, so it's all for the long haul.
Maybe the public could be asked to post the source code of the old development. Everyone would finish what they need for themselves.
Or, as a last resort, closed source code with no restrictions.
And when new version will come out with new endpoint and new realization, let it be commercial and then who needs the program will buy it.
:-) That would be unfair to those who bought it.
You better wait for the new version, hope it meets many needs, including yours :-)
:-) it wouldn't be fair to those who bought it.
Better wait for the new version, I hope it can satisfy many wishes, including yours :-).
Okay, I would agree and buy the old version, but preferably in the source code for improvement, and those who bought it, automatically open source.
Practical use of the old algorithm for you no longer exist, if a new algorithm, so let at least a normal programmer to learn from his code.
My ICQ: 38 zero zero 282 zero 3 for contact.
:-) it wouldn't be fair to those who bought it.
Better wait for the new version, I hope it will satisfy many wishes, including yours :-)
Igor, I think the new version will have to be made in two versions, commercial and free. I think a lot of people will want to buy the commercial version knowing what features it will have and what tasks it will allow to do :)
Dear Igor!
I wrote to you that "Working with the program shows that during the execution of complex macro programs, such as 7. From time to time the program hangs, which are expressed in the visible work of the script and actually no actions are not performed (in the logs nothing is written to the screen of the script does not change) in such a state the script can stay long enough. "As it turned out, this is due to hanging in the loop macro OneTest(), if for any reason the created copy of the tester is not running, the script will hang indefinitely. This can be overcome by including the code for forcing the tester to restart in the OneTest() macro program. For example, like this:
int j=0; // my
int a=0;
while(!rep){ //Wait for tester report to appear
j++;
rep = TimeOutReport(""); //Request for tester report to appear
if(IsStopped()){return(-1);} //If stopped by user - abort
SignExp = ExpertMessage("");
Comment("Time passed: "+TimeToStr(TimeLocal()-TimeStartExp,TIME_MINUTES|TIME_SECONDS)+" Expert Pass: "+SignExp);//display time
Sleep(1000*SpK); //wait
if(j>10){
a++;
StartTester(); //Start tester
TimeStartExp = TimeLocal(); //Write start time of optimization
if(StrToInteger(GlobVariable("Glob_BgColor"))!=0){col=Red;}else{col=Yellow;}
TextWrite(1, "ATTENTION! - TESTING !",col); //will display warning on the screen
j=0;
}
if (a>3) break;
}
I also wrote you that "When working with macro program" Complex"turned out that in the report file of this macro sometimes for some reason, the results of the profit obtained macro program different from the results obtained in manual mode with the same input data in the same tester! "This is due to incorrect reading in the macro program OneTest(), function FileTester() of the library WorkAnalyzer, data from the tester report file except for the first line of the tester report almost nothing else matches. Hence errors in profitability and other data. Since WorkAnalyzer library is closed I want the author to check and correct the problem. Studies have been performed on one currency pair and one timeframe. That excludes data distortions caused by averaging during multiple passes. I also noticed that the failure does not always occur. I have not found the system yet. It would be desirable to provide the source code to those users who have bought your product, to make changes and corrections.
Regards Sergey.